agent-provisioner
MANDATORY workflow for creating new agents. Enforces -01-agent-formatting standards from inception, ensuring every new agent is properly designed, researched, and curated before deployment.
npx skills add majiayu000/claude-skill-registry --skill agent-provisioner --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Creates and validates a new agent through a mandatory, end-to-end provisioning workflow. It requires domain analysis, tier classification, knowledge research, identity design, instruction authoring, tooling/MCP setup, and synthesis/validation, culminating in a fully curated agent definition and a curation session record.
How it works
- Activation: triggered via "/agent-provisioner" with parameters domain, name, and category, or by the prompt "I need a new agent for {description}".
- Step 1: DOMAIN ANALYSIS to determine domain, expertise areas, overlap with existing agents, and a recommended category; includes an interactive domain analysis interface and overlap detection using shell commands to search existing agents.
- Step 2: TIER CLASSIFICATION using the provided criteria to classify as Expert tier with associated token/instruction counts and model guidance, with options to confirm or adjust to Focused or PhD tiers.
- Step 3: KNOWLEDGE RESEARCH to identify authoritative sources, leverage methods like Firecrawl, and present each source for adjudication with a structured source-adjudication prompt.
- Step 4: IDENTITY DESIGN to define the interpretive lens and a vocabulary list (15-20 terms), with options to confirm or edit.
- Step 5: INSTRUCTION AUTHORING to generate tier-appropriate instructions, including an ALWAYS section (non-negotiables) and NEVER section (anti-patterns).
- Step 6: TOOLING & MCP to specify tool modes, MCP server selection, and proactive triggers.
- Step 7: SYNTHESIS & VALIDATION to produce the complete agent definition, validate against a CURATION-CHECKLIST, perform a human final review, and save to the appropriate category.
When to use it
Use when creating any new agent to ensure proper design, research, and curation before deployment. The workflow emphasizes domain justification, scoping, authoritative sourcing, and formalized instruction sets.
What it can touch
- Tools: claude-code (as declared in the skill).
- Outputs: fully curated agent definition, curation session record, agent registered in appropriate category.
Caveats
- Requires a human available for design decisions and domain knowledge for the new agent.
- The process is described as the only sanctioned method for dev-system agent creation. No outcomes are guaranteed beyond following the steps and producing the defined outputs.
# Agent Provisioner Skill ## Purpose **This is the ONLY sanctioned way to create new agents in dev-system.** Creating agents without this skill results in: - Missing knowledge source validation - Unverified tier classification - Incomplete MCP integration - Poor instruction quality - No curation record ## Activation ``` /agent-provisioner Parameters: domain: {domain-description} name: {proposed-agent-name} category: {parent-category in -03-agents} ``` Or simply: ``` I need a new agent for {description} ``` ## Workflow Overview ``` ┌─────────────────────────────────────────────────────────────────────────┐ │ AGENT PROVISIONING WORKFLOW │ ├─────────────────────────────────────────────────────────────────────────┤ │ │ │ Step 1: DOMAIN ANALYSIS │ │ ├─ What domain does this agent cover? │ │ ├─ What would a PhD in this domain know? │ │ └─ What existing agents overlap? (avoid duplication) │ │ ▼ │ │ Step 2: TIER CLASSIFICATION │ │ ├─ Focused (~500 tokens) - Bounded, clear scope │ │ ├─ Expert (~1500 tokens) - Specialized domain work │ │ └─ PhD (~3000 tokens) - Deep, novel challenges │ │ ▼ │ │ Step 3: KNOWLEDGE RESEARCH │ │ ├─ Identify authoritative sources (specs, official docs) │ │ ├─ Use
- Purpose
- Activation
- Workflow Overview
- Step 1: Domain Analysis
- Overlap Detection
- Step 2: Tier Classification
- Step 3: Knowledge Research
- Step 4: Identity Design
- Step 5: Instruction Authoring
- Step 6: Tooling & MCP
- Step 7: Synthesis & Validation
- Curation Record
- Signals
- Integration with Hooks
Search existing agents
grep -r "domain:" agents/-03-agents/ | grep -i "{domain-keywords}"
grep -r "proactive_triggers:" agents/-03-agents/ | grep -i "{trigger-patterns}"What does the agent-provisioner skill do?
MANDATORY workflow for creating new agents. Enforces -01-agent-formatting standards from inception, ensuring every new agent is properly designed, researched, and curated before deployment.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill agent-provisioner --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.